지난 며칠 동안 내 데비안 설치가 5~30분마다 무작위로 중단되었습니다. 화면이 정지됩니다. 가끔 GDM이 잠시 후 로그인 화면으로 돌아가는 경우가 있습니다. 다른 때에는 완전히 얼어붙을 때도 있습니다. 몇 주 만에 첫 번째 분산 업그레이드를 수행한 후에 시작되었습니다(Gnome 3.20으로 업그레이드 포함).
시스템 로그에서 찾은 내용은 다음과 같습니다. 매번 유사한 메시지 세트가 나타납니다.
Jun 17 11:18:10 piglet /usr/lib/gdm3/gdm-x-session[1160]: (EE) [mi] EQ overflowing. Additional events will be discarded until existing events are processed.
Jun 17 11:18:10 piglet /usr/lib/gdm3/gdm-x-session[1160]: (EE)
Jun 17 11:18:10 piglet /usr/lib/gdm3/gdm-x-session[1160]: (EE) Backtrace:
Jun 17 11:18:10 piglet /usr/lib/gdm3/gdm-x-session[1160]: (EE) 0: /usr/lib/xorg/Xorg (xorg_backtrace+0x4e) [0x55fd88968f6e]
Jun 17 11:18:10 piglet /usr/lib/gdm3/gdm-x-session[1160]: (EE) 1: /usr/lib/xorg/Xorg (mieqEnqueue+0x253) [0x55fd8894aa33]
Jun 17 11:18:10 piglet /usr/lib/gdm3/gdm-x-session[1160]: (EE) 2: /usr/lib/xorg/Xorg (QueuePointerEvents+0x52) [0x55fd88823632]
Jun 17 11:18:10 piglet /usr/lib/gdm3/gdm-x-session[1160]: (EE) 3: /usr/lib/xorg/Xorg (xf86PostMotionEvent+0xd6) [0x55fd8885a956]
Jun 17 11:18:10 piglet /usr/lib/gdm3/gdm-x-session[1160]: (EE) 4: /usr/lib/xorg/modules/input/synaptics_drv.so (0x7f0fbb634000+0x5f89) [0x7f0fbb639f$
Jun 17 11:18:10 piglet /usr/lib/gdm3/gdm-x-session[1160]: (EE) 5: /usr/lib/xorg/modules/input/synaptics_drv.so (0x7f0fbb634000+0x7532) [0x7f0fbb63b5$
Jun 17 11:18:10 piglet /usr/lib/gdm3/gdm-x-session[1160]: (EE) 6: /usr/lib/xorg/Xorg (0x55fd887b7000+0x940f8) [0x55fd8884b0f8]
Jun 17 11:18:10 piglet /usr/lib/gdm3/gdm-x-session[1160]: (EE) 7: /usr/lib/xorg/Xorg (0x55fd887b7000+0xb9392) [0x55fd88870392]
Jun 17 11:18:10 piglet /usr/lib/gdm3/gdm-x-session[1160]: (EE) 8: /lib/x86_64-linux-gnu/libc.so.6 (0x7f0fc3a6d000+0x334e0) [0x7f0fc3aa04e0]
Jun 17 11:18:10 piglet /usr/lib/gdm3/gdm-x-session[1160]: (EE) 9: /lib/x86_64-linux-gnu/libc.so.6 (ioctl+0x5) [0x7f0fc3b4e4f5]
Jun 17 11:18:10 piglet /usr/lib/gdm3/gdm-x-session[1160]: (EE) 10: /usr/lib/xorg/modules/drivers/intel_drv.so (0x7f0fbf3ef000+0x24f5d) [0x7f0fbf413f$
Jun 17 11:18:10 piglet /usr/lib/gdm3/gdm-x-session[1160]: (EE) 11: /usr/lib/xorg/modules/drivers/intel_drv.so (0x7f0fbf3ef000+0x28904) [0x7f0fbf4179$
Jun 17 11:18:10 piglet /usr/lib/gdm3/gdm-x-session[1160]: (EE) 12: /usr/lib/xorg/modules/drivers/intel_drv.so (0x7f0fbf3ef000+0x5c3ee) [0x7f0fbf44b3$
Jun 17 11:18:10 piglet /usr/lib/gdm3/gdm-x-session[1160]: (EE) 13: /usr/lib/xorg/Xorg (BlockHandler+0x4a) [0x55fd8880f5ba]
Jun 17 11:18:10 piglet /usr/lib/gdm3/gdm-x-session[1160]: (EE) 14: /usr/lib/xorg/Xorg (WaitForSomething+0x163) [0x55fd88965c33]
Jun 17 11:18:10 piglet /usr/lib/gdm3/gdm-x-session[1160]: (EE) 15: /usr/lib/xorg/Xorg (0x55fd887b7000+0x53a1e) [0x55fd8880aa1e]
Jun 17 11:18:10 piglet /usr/lib/gdm3/gdm-x-session[1160]: (EE) 16: /usr/lib/xorg/Xorg (0x55fd887b7000+0x57c03) [0x55fd8880ec03]
Jun 17 11:18:10 piglet /usr/lib/gdm3/gdm-x-session[1160]: (EE) 17: /lib/x86_64-linux-gnu/libc.so.6 (__libc_start_main+0xf0) [0x7f0fc3a8d5f0]
Jun 17 11:18:10 piglet /usr/lib/gdm3/gdm-x-session[1160]: (EE) 18: /usr/lib/xorg/Xorg (_start+0x29) [0x55fd887f8f99]
Jun 17 11:18:10 piglet /usr/lib/gdm3/gdm-x-session[1160]: (EE)
Jun 17 11:18:10 piglet /usr/lib/gdm3/gdm-x-session[1160]: (EE) [mi] These backtraces from mieqEnqueue may point to a culprit higher up the stack.
Jun 17 11:18:10 piglet /usr/lib/gdm3/gdm-x-session[1160]: (EE) [mi] mieq is *NOT* the cause. It is a victim.
Jun 17 11:18:10 piglet kernel: [ 1087.627729] [drm] stuck on render ring
Jun 17 11:18:10 piglet kernel: [ 1087.634979] [drm] GPU HANG: ecode 8:0:0xfffffffe, in Xorg [1162], reason: Ring hung, action: reset
Jun 17 11:18:10 piglet kernel: [ 1087.634995] [drm] GPU hangs can indicate a bug anywhere in the entire gfx stack, including userspace.
Jun 17 11:18:10 piglet kernel: [ 1087.635004] [drm] Please file a _new_ bug report on bugs.freedesktop.org against DRI -> DRM/Intel
Jun 17 11:18:10 piglet kernel: [ 1087.635012] [drm] drm/i915 developers can then reassign to the right component if it's not a kernel issue.
Jun 17 11:18:10 piglet kernel: [ 1087.635019] [drm] The gpu crash dump is required to analyze gpu hangs, so please always attach it.
Jun 17 11:18:10 piglet kernel: [ 1087.635029] [drm] GPU crash dump saved to /sys/class/drm/card0/error
Jun 17 11:18:10 piglet kernel: [ 1087.641887] drm/i915: Resetting chip after gpu hang
Jun 17 11:18:10 piglet /usr/lib/gdm3/gdm-x-session[1160]: [mi] Increasing EQ size to 1024 to prevent dropped events.
Jun 17 11:18:10 piglet /usr/lib/gdm3/gdm-x-session[1160]: [mi] EQ processing has resumed after 37 dropped events.
Jun 17 11:18:10 piglet /usr/lib/gdm3/gdm-x-session[1160]: [mi] This may be caused by a misbehaving driver monopolizing the server's resources.
Jun 17 11:18:13 piglet pulseaudio[916]: [pulseaudio] sink-input.c: Failed to create sink input: sink is suspended.
Jun 17 11:18:18 piglet kernel: [ 1095.635999] [drm] stuck on render ring
Jun 17 11:18:18 piglet kernel: [ 1095.642984] [drm] GPU HANG: ecode 8:0:0xfffffffe, in Xorg [1162], reason: Ring hung, action: reset
Jun 17 11:18:18 piglet kernel: [ 1095.643414] [drm:i915_set_reset_status [i915]] *ERROR* gpu hanging too fast, banning!
Jun 17 11:18:18 piglet kernel: [ 1095.646268] drm/i915: Resetting chip after gpu hang
Jun 17 11:18:18 piglet /usr/lib/gdm3/gdm-x-session[1160]: (EE) intel(0): Failed to submit rendering commands (Input/output error), disabling acceleration.
내 시스템은 Debian Stretch에서 완전히 최신 상태입니다.
어떻게 하면 이를 방지할 수 있나요? Intel i915 드라이버에서 드라이버를 전환할 수 있습니까? 아니면 grub 구성에서 DRM이나 다른 요소를 비활성화하시겠습니까?
답변1
로그는 취해야 할 조치를 알려줍니다.
GPU 정지는 사용자 공간을 포함하여 전체 gfx 스택의 어디에서나 버그를 나타낼 수 있습니다.
제출해 주세요새로운bugs.freedesktop.org의 DRI에 대한 버그 보고서 -> DRM/Intel drm/i915 커널 문제가 아닌 경우 개발자는 올바른 구성 요소에 재배포할 수 있습니다.
GPU 정지를 분석하려면 GPU 크래시 덤프가 필요하므로 항상 연결하세요. GPU 크래시 덤프는 다음 위치에 저장됩니다.
/sys/class/drm/card0/error
(알려진 버그이고 이미 어딘가에서 수정되지 않은 경우) 이렇게 하면 문제가 즉시 해결되지는 않지만 문제를 해결하는 데 도움이 될 것입니다. 개발자의 i915반응은 꽤 과격했습니다.